COVID-19 vaccine access inequity was a major challenge during the pandemic. This inequity was present between countries and regions and within cities. We developed a novel approach to measure and improve vaccine access equity to address this issue. Our approach first created a vaccination attainment index based on CDC COVID-19 data. We then selected the most relevant spatial, e. g., the density of medical facilities, and socioeconomic factors, to train an XGBoost model on the county level of the United States. Using this model on census tracts within counties, we used the Gini and Theil indices to measure equity. We identified the main drivers of vaccine access based on SHAP values. With the main drivers identified (percentage of American Indian and Alaska native population, health insurance coverage, and transportation options), we conducted a case study on Cambridge, MA. We improved the short-term access equity by adjusting each census tract’s density of medical facilities (from Gini 0.14 to 0.13). Our novel approach provides decision-makers with a tool to identify and address drivers of vaccine access equity in their region and predict vaccination attainment on the tract level. These insights are crucial to ensuring equal access to vaccines and other essential healthcare services for everyone.